CN114338957B - Video denoising method and system - Google Patents
Video denoising method and system Download PDFInfo
- Publication number
- CN114338957B CN114338957B CN202210245613.6A CN202210245613A CN114338957B CN 114338957 B CN114338957 B CN 114338957B CN 202210245613 A CN202210245613 A CN 202210245613A CN 114338957 B CN114338957 B CN 114338957B
- Authority
- CN
- China
- Prior art keywords
- slice
- key
- value pair
- key value
- conversion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000010276 construction Methods 0.000 claims abstract description 37
- 230000009467 reduction Effects 0.000 claims abstract description 34
- 238000006243 chemical reaction Methods 0.000 claims description 37
- 230000009466 transformation Effects 0.000 claims description 32
- 230000008707 rearrangement Effects 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 8
- WYROLENTHWJFLR-ACLDMZEESA-N queuine Chemical compound C1=2C(=O)NC(N)=NC=2NC=C1CN[C@H]1C=C[C@H](O)[C@@H]1O WYROLENTHWJFLR-ACLDMZEESA-N 0.000 claims description 8
- 230000001186 cumulative effect Effects 0.000 claims description 7
- 238000003672 processing method Methods 0.000 claims 1
- 230000036961 partial effect Effects 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000035508 accumulation Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000002829 reductive effect Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Landscapes
- Image Processing (AREA)
- Picture Signal Circuits (AREA)
Abstract
The invention discloses a video denoising method and a system, wherein the method comprises the following steps: acquiring a video frame, pre-establishing a first slice construction table, generating a first slice of the video frame according to the first slice construction table, and generating a first slice key value pair; pre-establishing a second slice construction table, generating a second slice of the video frame according to the second slice construction table, and generating a second slice key value pair, wherein the second slice comprises partial data of the first slice; performing first noise reduction processing on the second slice according to the second slice key value and a Mapper class to generate a third slice key value pair; performing second noise reduction processing on the third slice key value pair according to Reduce class to generate a fourth slice key value pair; and performing inverse solution on the fourth slice key value pair to generate a fifth slice key value pair, and restoring to generate the noise reduction video stream at the display end according to the fifth slice key value pair.
Description
Technical Field
The invention relates to the technical field of videos, in particular to a video denoising method and system.
Background
Among the prior art traditional time-space domain noise reduction generally adopts time domain processing to the video static area, adopts space domain processing to the video motion area, and above-mentioned prior art has the defect that: the motion detection algorithm needs to perform motion compensation on a background difference model, and the motion region spatial processing is weaker in random noise processing and is easy to bring motion blur.
Disclosure of Invention
One of the purposes of the invention is to provide a video noise reduction method and system, the method and system utilize a MapReduce system to perform noise reduction processing on videos without dividing the videos into motion or static, and utilize Map task and Reduce task to realize video noise reduction, so that the noise reduction effect on motion areas is obvious.
Another objective of the present invention is to provide a method and a system for video noise reduction, which can reduce the detection process of a motion region, avoid the smear of a video caused by misjudgment of motion, and simultaneously process the static and motion noise of the video consistently, so as to achieve a better effect in the environment of cross-bottom signal-to-noise ratio.
Another object of the present invention is to provide a method and a system for reducing noise in a video, where the method and the system respectively perform noise reduction processing by constructing a first slice and a second slice and generating corresponding key value pairs, and a traversal construction is performed in the first slice and the second slice construction process while considering spatial correlations such as a line direction, a longitudinal direction, a connected domain, a z-type, and the like, so that a video does not need to be divided into a still state and a motion, thereby improving the effect of reducing noise in the video.
To achieve at least one of the above objects, the present invention further provides a video denoising method, comprising the steps of:
acquiring a video frame, pre-establishing a first slice construction table, generating a first slice of the video frame according to the first slice construction table, and generating a first slice key value pair;
pre-establishing a second slice construction table, generating a second slice of the video frame according to the second slice construction table, and generating a second slice key value pair, wherein the second slice comprises partial data of the first slice;
performing first noise reduction processing on the second slice according to the second slice key value and a Mapper class to generate a third slice key value pair;
performing second noise reduction processing on the third slice key value pair according to Reduce class to generate a fourth slice key value pair;
and performing inverse solution on the fourth slice key value pair to generate a fifth slice key value pair, and restoring to generate the noise reduction video stream at the display end according to the fifth slice key value pair.
According to a preferred embodiment of the present invention, the method for constructing the first slice includes: and traversing with any spatial correlation comprising a row direction, a column direction, a connected domain scanning type or a Z type according to a pre-established first slice configuration table to generate the first slice.
According to another preferred embodiment of the inventionIn an embodiment, the method for constructing the first slice key-value pair includes: acquiring a first video data point of a video frame, using the coordinate of the first video data point as a first auxiliary information kref1 value, using the first auxiliary information kref1 as a key name, constructing a key value vref1 of a first key-value pair according to a traversed video data point sequence, and further generating the first key-value pair according to the first auxiliary information kref1 and the key value vref1<k 1 ,v 1 >。
According to another preferred embodiment of the present invention, the method for constructing the second slice key-value pair includes: and combining the addresses of the first and second slices into a second auxiliary information kref2 value and using the second auxiliary information kref2 value as the key name of the second key-value pair, taking the corresponding data values according to the traversal order of the first slice to construct a key value vref2 of the second key-value pair, and further constructing the second key-value pair < k, v > according to the second auxiliary information kref2 and the key value vref 2.
According to another preferred embodiment of the present invention, the method for constructing the third slice key-value pair includes: sequentially multiplying the key values v of the second slice key value pair with the corresponding transformation bases according to the lengths of the transformation bases to obtain a first transformation base data processing result v Tx Processing the result v according to the first transformation base data Tx Carrying out rearrangement processing, wherein the rearrangement processing comprises merging the first conversion base data processing results of different key value v data segments to obtain a second conversion and data processing result v Tx2 Processing the second transformation base data result v Tx2 Multiplying the conversion bases in sequence according to the lengths of the conversion bases to obtain a third conversion base data processing result v of the second slice T 。
According to another preferred embodiment of the present invention, the method for constructing the third slice key-value pair further comprises: a third transform base data processing result v obtained from the third transform base data processing result according to a transform characteristic of the transform base T Extracting a specific number of data as k T Wherein said k is T And said third transform base data processing result v T Make up key-value pairs<k T ,v T >To said k T Setting the bit width of the key, and defaulting the processing mode exceeding the corresponding bit width to be upper limit truncation.
According to another preferred embodiment of the present invention, an intra-group threshold and an intra-group number upper-limit threshold of a third slice key-value pair are set, defining said third slice key-value pair<km,vm>Wherein a key name value range in the third slice key value pair is km ═ km0, kmid ]In which
Wherein kvthr is an intra-group threshold, kmid is a refinement id of the third slice; wherein vm is [ k, v ═ v T ,count]Where count is the number of second slices, k is the second slice key name, V T The count is the maximum number of key-value pairs in the third slice group designed for the transform result corresponding to the second slice key-value.
According to another preferred embodiment of the invention, the method of constructing a fourth slice from said third slice comprises the steps of: according to the third slice key-value pair<km,vm>Extracting count second slice construction key value pairs<k,v T >Setting the intensity noise reduction percentage Nper, and calculating a key value v T Sum of absolute values vTsum and calculate the cumulative sum, when the cumulative sum is greater than or equal to (Nper x vTsum), v at that time is T As v Tn Is constructed by subjecting v to Tn Obtaining a first transformation base data processing result v of a fourth slice according to the multiplication summation of the transformation base Q Qx And processing the result v according to the first transform base data of the fourth slice Qx Obtaining a second transformation base data processing result v of the fourth slice after rearrangement Qx2 And according to the second conversion base data processing result of the fourth slice, multiplying the conversion base data processing result of the fourth slice by each conversion base in sequence according to the length of the conversion base Q to obtain a third conversion base data processing result v of the fourth slice Q Wherein a third transform base data processing result v of the fourth slice Q A key value vr ═ v for the fourth slice Q And the fourth slice has the key name kr ═ k.
To achieve at least one of the above objects, the present invention further provides a video denoising system, which performs the above video denoising method.
The present invention further provides a computer-readable storage medium having stored thereon a computer program executable by a processor for performing the method of video noise reduction.
Drawings
Fig. 1 is a schematic flow chart of a video denoising method according to the present invention.
FIG. 2 shows a schematic of an input system of the present invention.
Fig. 3 shows a schematic view of a noise reduction system according to the invention.
Fig. 4 shows a schematic diagram of an output system according to the present invention.
Fig. 5 shows a schematic view of the configuration of the first and second slices of the present invention.
FIG. 6 shows a schematic view of a backward scan of the first and second slice configurations and a table of their configurations in accordance with the present invention.
FIG. 7 is a schematic view of a Z-scan of a first slice and a second slice after their construction according to the present invention and a table showing the construction thereof.
FIG. 8 shows v in the present invention T The structure is schematic.
Detailed Description
The following description is presented to disclose the invention so as to enable any person skilled in the art to practice the invention. The preferred embodiments in the following description are given by way of example only, and other obvious variations will occur to those skilled in the art. The basic principles of the invention, as defined in the following description, may be applied to other embodiments, variations, modifications, equivalents, and other technical solutions without departing from the spirit and scope of the invention.
It is understood that the terms "a" and "an" should be interpreted as meaning that a number of one element or element is one in one embodiment, while a number of other elements is one in another embodiment, and the terms "a" and "an" should not be interpreted as limiting the number.
Referring to fig. 1 to 8, the present invention discloses a method and a system for video denoising, wherein the method mainly comprises the following steps: first, an image of each frame of a video is acquired, and a first slice construction table and a second slice construction table are constructed, wherein the first slice construction table generates video slice data with a first slice size of 8 × 8 or 16 × 16, and the second slice construction table generates a second slice containing the first slice image data with a size of 2 × 2 or 4 × 4. It should be noted that the sizes of the first slice and the second slice may be set according to the video frame size, and the specific size of the slice is not particularly limited by the present invention. After the first slice and the second slice are constructed, further generating key value pairs of the first slice and the second slice, further generating a third slice key value pair according to the key value pair of the second slice, denoising the third slice key value pair to generate a denoised fourth slice key value pair, reversely resolving the fourth slice key value pair according to a second slice construction table to generate a fifth slice key value pair, and further reversely resolving the fifth slice key value pair according to the first slice construction table to output a corresponding denoised video stream.
In particular, the video denoising requires borrowing a MapReduce system, where the MapReduce system can identify corresponding key-value pairs generated from the first and second slices. Wherein the method of constructing the first slice comprises: pre-establishing a first slice configuration table, scanning the video frame according to the first slice configuration table in any one manner including, but not limited to, a row direction, a column direction, a connected domain, and a z-type to generate corresponding first slice data, wherein the first slice configuration method may further include: inputting the gray data, RGB data and yuv data of the video frame, so that the first slice can be suitable for processing the video data with different formats. Further, the first slice configuration table may generate a vref1 value of the first slice according to a first video data point in the first slice as an auxiliary information kref1 value, and further according to a traversal mode and a traversal order, wherein the vref1 value represents position data of the first slice, the auxiliary information is a first data coordinate of the slice configuration, and according to the auxiliary information kref1 value and vref1, constructing key-value pairs for the first slice<k 1 ,v 1 >. And saving the key value pair to an Hbase system, and reversely solving by using a standby output system. Wherein the construction of the first slice may further comprise repeated video frames.
The second slice construction method comprises the following steps: and a second slice construction table is constructed in advance, the video frame is scanned according to any one mode including but not limited to a row direction, a column direction, a connected domain and a z type according to the second enterprise slice construction table to generate corresponding second slice data, a first slice construction sequence contained in the second slice is used as auxiliary information kref2 of the second slice, a vref1 value of the second slice is further generated according to the scanning mode and the scanning sequence of the second slice, a key < k, v > of the second slice is constructed according to the auxiliary information kref2 of the second slice and the vref1 value of the second slice, the key value pair < k, v > of the second slice is stored in an Hbase database, and a standby output system reversely solves the key value pair < k, v > of the second slice. That is, the MapReduce system inputs data for the key-value pair < k, v > of the second slice construction result. The second slice construction method further comprises: gray data, RGB data, and yuv data of the video frame are input to adapt the second slice to video data of different formats, and the second slice data may contain repeated first slice data.
Generating key-value pairs of the second slice <k,v>Then, according to the second slice key value pair<k,v>Inputting the signal into a noise reduction system for noise reduction processing, wherein the method for noise reduction processing comprises the following steps: for the second slice key-value pair of input<k,v>Performing first noise reduction processing by using a Mapper class, wherein the first noise reduction processing comprises the following steps: for the second slice key value pair<k,v>The Mapper class construction is carried out again to generate a third slice key value pair<km,vm>Wherein the Mapper class constructs a third slice key-value pair<km,vm>Including a pair of inputs<k,v>The key value v is subjected to data processing to obtain v T Wherein the format composition of the output value vm of the third slice key value pair Mapper class comprises an original key name k and the data processing result v T SaidMapper based third slice key-value pairs<km,vm>Further comprises updating the input second slice key value k to obtain k T . Wherein updating the second slice key value k to obtain k T The method comprises processing the data to obtain v T Sorting according to the information attention degree, wherein the information attention degree is the importance degree of the transformation base characteristics, the bit width is more before the high ranking, the respective key width is set according to the information attention degree, and the third slice key value pair<km,vm>The update includes an add merge operation, wherein the add merge operation is included in an existing vT i In which a newly added vT is added j 。
The invention further provides for said third slice key-value pair<km,vm>Performing a second denoising operation using Reduce class, wherein the second denoising operation comprises: pair the third slice key-value pair by Reduce class<km,vm>Reconstructing key-value pairs<k r ,v r >Wherein the key-value pair<k r ,v r >Is a fourth slice key-value pair, said fourth slice key-value pair<k r ,v r >The construction method comprises the following steps: third slice key-value pair input by Reduce class pair<km,vm>The vm value of (a) is subjected to data processing to obtain v Tr . And extracting corresponding v for vm by Reduce class computation Ti Data, go on v Ti Inter-data processing to obtain v Tr (ii) a Extracting corresponding k for vm i Data, wherein i ranges from 1 to vnummax; where vnummax is the second number of slices contained within the corresponding third slice construction group, in terms of k i Recovery of kr, vTr i Restoring vr;<kr,vr>the updating also includes a merging process, i.e. after an existing vr i Internally merging vr according to merging rules j . Both the first noise reduction operation and the second noise reduction operation can be executed through Map task and Reduce task, and on the basis of the prior art, the principles of how to recover and combine processing are not described in detail, and a person skilled in the art can implement the operations according to the data processing requirements.
And when the fourth slice key value pair < kr, vr > is obtained, performing inverse solution on the fourth slice key value pair < kr, vr > according to the second slice configuration table, wherein the inverse solution data is fifth slice data, the size of the fifth slice is the same as that of the second slice, and the output system further performs inverse solution on the fifth slice data according to the first slice configuration table to generate a noise-reduced video stream. If the fifth slice data has a repeated structure, the noise reduction system may further perform solution and reconstruction operations using processes including, but not limited to, mean processing and gaussian processing, and how to perform solution and reconstruction is not described in detail in the present invention.
In order to better illustrate the technical effects of the present invention, the present invention provides the following specific embodiments to illustrate the above technical process:
referring to fig. 5, the upper left square indicates input video data as a whole, different numbers indicate different partition data thereof, the upper right square indicates output first slice data, and the lower side indicates second slice data; finally, the second slice data can adopt a Text input format to carry out < k, v > construction, namely the line address offset is k, and the line data is v; in the present embodiment, as shown in fig. 6, a video frame is scanned in a size of 8 × 8 to construct a first slice data of len1 — 64 pixels; as shown in fig. 7, scanning a video frame with a size of 4 × 4 constructs second slice data with l en2 ═ 64 × 16 ═ 1024 pixels, where the scanned index is corresponding to the first and second slice construction tables, which is generally different from the slice data storage, i.e. its key values < kref, vref > are redistributed and calculated by the hbase system.
Further, constructing a third slice key-value pair for the v value of the second slice, first constructing a transformation base T: T1-T8:
T1＝[362,362,362,362,362,362,362,362]；
T2＝[225,460,460,225,-225,-460,-460,-225]；
T3＝[583,412,-412,-583,-86,86,-86,86]；
T4＝[-86,86,-86,86,583,412,-412,-583]；
T5＝[724,-724,0,0,0,0,0,0]；
T6＝[0,0,724,-724,0,0,0,0]；
T7＝[0,0,0,0,724,-724,0,0]；
T8＝[0,0,0,0,0,0,724,-724]；
the length, number and value of the transformation base are all determined by the requirement of the noise reduction system corresponding to the size of the first slice, and the invention is only illustrated by way of example, and the transformation base can be changed arbitrarily according to the requirement.
For the second slice data, according to the key value v in the second slice data, sequentially multiplying and summing each transformation base according to the length of the transformation base to obtain a processing result v of the first transformation base data Tx Processing the result v according to the first transformation base data Tx Carrying out rearrangement processing, wherein the rearrangement processing comprises merging the first conversion base data processing results of different key value v data segments to obtain a second conversion and data processing result v Tx2 Processing the second transform base data result v Tx2 Multiplying the conversion bases in sequence according to the lengths of the conversion bases to obtain a third conversion base data processing result v of the second slice T . Reference numeral 7xy shown in fig. 7 represents the y-th data type corresponding to the x-th group of operation data, respectively.
Further processing the result v according to the third transformation base data T Proceed to k T The transformation characteristics of the transformation bases T1-T8 in this example, 5 data of 1 st, 2 nd, 3 th, 9 th, 17 th, etc. in vT are taken as k T Is composed of sequentially arranged keys with width of [32,10,6 ] respectively]bit, and the default processing mode exceeding the corresponding bit width is upper limit truncation, wherein examples of the upper limit truncation operation include: according to v T The key bit width 32 bit set to the 1 st data of (1), that is, the maximum value that can be expressed is 2 31 -1, that actual data may exceed 2 31 1, then the direct truncation is noted as 2 31 And (4) obtaining the product with the formula of-1. Further according to key value pair<k T ,v T >Structure of the device<km,vm>Wherein a key name value range in the third slice key value pair is km ═ km0, kmid]Wherein
Where kvthr is the intra-group threshold and kmid is the refinement id of the third slice; wherein vm is [ k, v ═ v T ,count]Where count is the number of second slices and k is the second slice key name，V T The count is the maximum number of key-value pairs in the third slice group designed for the transform result corresponding to the second slice key-value. If count is greater than kvnumthr, it is current<km,vm>No longer updated, kmid increments the cycle.
The method of constructing a fourth slice key-value pair from the third slice key-value pair comprises: setting a transformation base Q:
Q1＝[362,362,512,0,724,-124,0,124]；
Q2＝[362,362,512,0,-724,-124,0,124]；
Q3＝[362,362,-512,0,124,724,-124,0]；
Q4＝[362,362,-512,0,124,-724,-124,0]；
Q5＝[362,-362,0,512,0,124,724,-124]；
Q6＝[362,-362,0,512,0,124,-724,-124]；
Q7＝[362,-362,0,-512,-124,0,124,724]；
Q8＝[362,-362,0,-512,-124,0,124,-724]；
according to the third slice key-value pair<km,vm>Extracting count second slice construction key value pairs<k,v T >Setting the intensity noise reduction percentage Nper, for example, the intensity noise reduction percentage may be set to 80, and calculating the key value v T Sum of absolute values vTsum and calculate the cumulative sum, when the cumulative sum is greater than or equal to (Nper x vTsum), v at that time is T As v Tn (ii) structure v tn Representing selected data in accordance with noise reduction ratio, denoted v T Sequentially accumulating, if the sum of the accumulations is greater than or equal to (Nper v) Tsum ) Then v follows T Set to 0, no subsequent fourth slice construction is required, and the cumulative sum is less than (Nper x v) Tsum ) V is t Is denoted by v tn (more than one, meaning that all v within this range are satisfied T ) To v is to v tn V is obtained by performing the fourth slice structure, i.e. performing Q transformation base for several times tr . The conversion base T and the conversion base Q are reciprocal conversion bases.
Will v is Tn Multiplying and summing the conversion base Q length with the corresponding conversion base in sequence to obtain the fourth sliceA result v of processing the transformed base data Qx And processing the result v according to the first transform base data of the fourth slice Qx Obtaining a second transformation base data processing result v of the fourth slice after rearrangement Qx2 。
Processing the result v further from a second transform base data of the fourth slice Qx2 Multiplying the conversion bases in sequence according to the length of the conversion base Q to obtain a third conversion base data processing result v of the fourth slice Q Wherein the third transform base data of the fourth slice is processed into a result v Q As the key value vr of the fourth slice, and the key name kr ═ k of the fourth slice, the fourth slice key value pair can be constructed<kr,vr>. It is worth mentioning that the two transformation bases in the present invention are only examples.
That is to say, the structure of the third slice is implemented in the mapper class of the denoising system, the structure of the fourth slice is implemented in the Reduce class of the denoising system, the fourth slice key value pair of the structure is further solved reversely according to the second slice construction table to obtain fifth slice data, and the fifth slice data is further solved reversely according to the first slice construction table to obtain a final denoising video stream. The present invention does not improve the inverse solution operation, which is to solve the video stream reversely only according to the first two-slice construction table, and belongs to the conventional technical means in the field.
In particular, according to embodiments of the present disclosure, the processes described above with reference to the flow diagrams may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication section, and/or installed from a removable medium. The computer program, when executed by a Central Processing Unit (CPU), performs the above-described functions defined in the method of the present application. It should be noted that the computer readable medium mentioned above in the present application may be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wire segments, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present application, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In this application, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless section, wire section, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
It will be understood by those skilled in the art that the embodiments of the present invention described above and illustrated in the drawings are given by way of example only and not by way of limitation, the objects of the invention having been fully and effectively achieved, the functional and structural principles of the present invention having been shown and described in the embodiments, and that various changes or modifications may be made in the embodiments of the present invention without departing from such principles.
Claims (4)
1. A method for video denoising, the method comprising the steps of:
generating a vref1 value of the first slice according to a traversal mode and a traversal order by taking a first video data point coordinate in the first slice as auxiliary information kref1 value, wherein the vref1 value represents position data of the first slice, and constructing a key value pair < k1, v1> of the first slice according to the auxiliary information kref1 value and vref 1;
pre-establishing a second slice configuration table, combining the addresses of the first and second slices into auxiliary information kref2 of the second slice according to the second slice configuration table, using the auxiliary information kref2 as the key name of the key value pair of the second slice, taking the corresponding data values according to the traversal sequence of the first slice to construct a key value vref2 of the key value pair of the second slice, and further constructing a second slice key value pair < k, v > according to the auxiliary information kref2 and the key value vref2 of the second slice;
performing first denoising processing on the second slice according to the Mapper class according to the second slice key value pair to generate a third slice key value pair, wherein the first denoising processing comprises:
Constructing a third slice key-value pair: sequentially multiplying the key values v of the second slice key value pair with the corresponding transformation bases T according to the length of the transformation bases to obtain a first transformation base data processing result v Tx Processing the result v according to the first transformation base data Tx Carrying out rearrangement processing, wherein the rearrangement processing comprises merging the first conversion base data processing results of different key value v data segments to obtain a second conversion base data processing result v Tx2 Processing the second transformation base data result v Tx2 Multiplying the conversion bases in sequence according to the lengths of the conversion bases to obtain a third conversion base data processing result v of the second slice T ；
Processing the result v from a third transform base data according to the transform characteristics of said transform base T Extracting a specific number of data as k T Wherein said k is T And said third transform base data processing result v T Make up key-value pairs<k T , v T >To said k T Setting the bit width of the key, and defaulting the processing mode exceeding the corresponding bit width as upper limit truncation;
setting an intra-group threshold and an intra-group number upper limit threshold of a third slice key-value pair, defining the third slice key-value pair<km,vm>Wherein a key name value range in the third slice key value pair is km = [ km0, kmid]Wherein 
Where kvthr is the intra-group threshold and kmid is the refinement id of the third slice; wherein vm = [ k, v = T ,count]K is the name of the second slice key, V T The count is the maximum number of key value pairs in the third designed slice group as the conversion result corresponding to the second slice key value;
and carrying out second noise reduction processing on the third slice key value pair according to Reduce classes to generate a fourth slice key value pair, wherein the second noise reduction processing method comprises the following steps:
constructing a fourth slice from the third sliceThe slicing method comprises the following steps: according to the third slice key-value pair<km,vm>Extracting count second slice construction key value pairs<k,v T >Setting the intensity noise reduction percentage Nper, and calculating a key value v T Sum of absolute values v Tsum And calculating the cumulative sum when the cumulative sum is greater than or equal to (Nper v) Tsum ) When v is reached, v at that time is measured T As v Tn Is constructed by subjecting v to Tn Multiplying and summing the conversion base length and the corresponding conversion base Q in sequence to obtain a first conversion base data processing result v of the fourth slice Qx And processing the result v according to the first transform base data of the fourth slice Qx Obtaining a second transformation base data processing result v of the fourth slice after rearrangement Qx2 And according to the second conversion base data processing result of the fourth slice, multiplying the conversion base data processing result of the fourth slice by each conversion base in sequence according to the length of the conversion base Q to obtain a third conversion base data processing result v of the fourth slice Q Wherein a third transform base data processing result v of the fourth slice Q The key value vr = v for the fourth slice Q And the fourth slice has a key name kr = k, resulting in a fourth slice key-value pair<kr,vr>Wherein the T transformation base and the Q transformation base are positive and negative transformation bases;
and performing inverse solution on the fourth slice key value pair according to the second slice configuration table to obtain fifth slice data, and further performing inverse solution on the fifth slice data according to the first slice configuration table to obtain a final noise reduction video stream.
2. A method for video denoising as claimed in claim 1, wherein the method for constructing the first slice comprises: and traversing with any spatial correlation comprising a row direction, a column direction, a connected domain scanning type or a Z type according to a pre-established first slice configuration table to generate the first slice.
3. A video noise reduction system, characterized in that said system performs a video noise reduction method according to any of claims 1-2.
4. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program which can be executed by a processor to perform a video denoising method according to any one of claims 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210245613.6A CN114338957B (en) | 2022-03-14 | 2022-03-14 | Video denoising method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210245613.6A CN114338957B (en) | 2022-03-14 | 2022-03-14 | Video denoising method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114338957A CN114338957A (en) | 2022-04-12 |
| CN114338957B true CN114338957B (en) | 2022-07-29 |
Family
ID=81033919
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210245613.6A Active CN114338957B (en) | 2022-03-14 | 2022-03-14 | Video denoising method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114338957B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110633733A (en) * | 2019-08-14 | 2019-12-31 | 中国平安财产保险股份有限公司 | Intelligent image matching method and device and computer readable storage medium |
| CN112837248A (en) * | 2021-04-22 | 2021-05-25 | 杭州雄迈集成电路技术股份有限公司 | Video stream denoising method and system |
| CN112991235A (en) * | 2021-05-18 | 2021-06-18 | 杭州雄迈集成电路技术股份有限公司 | Video noise reduction method and video noise reduction terminal |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8799228B2 (en) * | 2011-06-29 | 2014-08-05 | Nokia Corporation | Method and apparatus for providing a list-based interface to key-value stores |
| US8700683B2 (en) * | 2011-10-24 | 2014-04-15 | Nokia Corporation | Method and apparatus for providing a key-value based storage interface |
| US9564175B2 (en) * | 2013-04-02 | 2017-02-07 | International Business Machines Corporation | Clustering crowdsourced videos by line-of-sight |
| US10902468B2 (en) * | 2014-06-23 | 2021-01-26 | Board Of Regents, The University Of Texas System | Real-time, stream data information integration and analytics system |
| US10180969B2 (en) * | 2017-03-22 | 2019-01-15 | Www.Trustscience.Com Inc. | Entity resolution and identity management in big, noisy, and/or unstructured data |
| US11443416B2 (en) * | 2019-08-30 | 2022-09-13 | Sas Institute Inc. | Techniques for image content extraction |
-
2022
- 2022-03-14 CN CN202210245613.6A patent/CN114338957B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110633733A (en) * | 2019-08-14 | 2019-12-31 | 中国平安财产保险股份有限公司 | Intelligent image matching method and device and computer readable storage medium |
| CN112837248A (en) * | 2021-04-22 | 2021-05-25 | 杭州雄迈集成电路技术股份有限公司 | Video stream denoising method and system |
| CN112991235A (en) * | 2021-05-18 | 2021-06-18 | 杭州雄迈集成电路技术股份有限公司 | Video noise reduction method and video noise reduction terminal |
Non-Patent Citations (2)
| Title |
|---|
| 基于MapReduce的公路视频图像车型分类研究;许晓珑等;《电视技术》;20160317(第03期);全文 * |
| 基于大数据的公路监控智能视频分析系统的构建;靳明等;《电子设计工程》;20200220(第04期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114338957A (en) | 2022-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chandel et al. | Image filtering algorithms and techniques: A review | |
| He et al. | Guided image filtering | |
| CN110189260B (en) | Image noise reduction method based on multi-scale parallel gated neural network | |
| Gu et al. | A single image dehazing method using average saturation prior | |
| KR20100112162A (en) | Methods for fast and memory efficient implementation of transforms | |
| Park | 2D discrete Fourier transform on sliding windows | |
| Foare et al. | Semi-linearized proximal alternating minimization for a discrete Mumford–Shah model | |
| Majeeth et al. | Gaussian noise removal in an image using fast guided filter and its method noise thresholding in medical healthcare application | |
| CN113066034A (en) | Face image restoration method and device, restoration model, medium and equipment | |
| Zhang et al. | Image denoising by using PDE and GCV in tetrolet transform domain | |
| Kazar et al. | Evolutionary Cellular Automata for Image Segmentation and Noise Filtering Using Genetic Algorithms. | |
| Willett | Multiscale analysis of photon-limited astronomical images | |
| CN115131229A (en) | Image noise reduction and filtering data processing method and device and computer equipment | |
| Jin et al. | A new poisson noise filter based on weights optimization | |
| CN114338957B (en) | Video denoising method and system | |
| Han et al. | Multiplicative noise removal combining a total variation regularizer and a nonconvex regularizer | |
| WO2024082602A1 (en) | End-to-end visual odometry method and apparatus | |
| KR102135459B1 (en) | Denoising method in image processing device and method for generating feature therefor | |
| Huang et al. | Efficient and Effective Total Variation Image Super‐Resolution: A Preconditioned Operator Splitting Approach | |
| Yüksel et al. | Performance enhancement of image impulse noise filters by image rotation and fuzzy processing | |
| Wang et al. | Fast image inpainting using exponential-threshold POCS plus conjugate gradient | |
| He et al. | Adaptive shearlet-regularized image deblurring via alternating direction method | |
| Peng et al. | Detail enhancement for infrared images based on propagated image filter | |
| Li | Example-based learning using heuristic orthogonal matching pursuit teaching mechanism with auxiliary coefficient representation for the problem of de-fencing and its affiliated applications | |
| Sivaramakrishnan et al. | Universal denoising of discrete-time continuous-amplitude signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right |
Denomination of invention: A method and system of video noise reduction Effective date of registration: 20221211 Granted publication date: 20220729 Pledgee: Hangzhou Fuyang Sub branch of Zheshang Bank Co.,Ltd. Pledgor: Hangzhou xiongmai integrated circuit technology Co.,Ltd. Registration number: Y2022330003571 |
|
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 311422 4th floor, building 9, Yinhu innovation center, 9 Fuxian Road, Yinhu street, Fuyang District, Hangzhou City, Zhejiang Province Patentee after: Zhejiang Xinmai Microelectronics Co.,Ltd. Address before: 311422 4th floor, building 9, Yinhu innovation center, 9 Fuxian Road, Yinhu street, Fuyang District, Hangzhou City, Zhejiang Province Patentee before: Hangzhou xiongmai integrated circuit technology Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |